Math environment flanking rules

[notriddle]

The math extension from this repository is pretty eager.

When I reimplemented these rules in pulldown-cmark and tested them against crates.io, a lot of text that wasn't intended to be inline math got matched. For example, this snippet:

You can get a const reference to any layer or a mutable reference to the top-most layer using
methods that match ref_$fieldname() or mut_$fieldname(). If you need references to multiple
layers simultaneously, you can call the view() method which will return a ViewStruct that
contains public fields referencing each layer [^0].

This error reproduces in pandoc running in gfm mode, with an error "Could not convert TeX math fieldname() or mut_".

Actual Github-Flavored Markdown parses it the way the author intended, and so does VSCode's markdown-it plugin:

You can get a const reference to any layer or a mutable reference to the top-most layer using
methods that match ref_$fieldname() or mut_$fieldname(). If you need references to multiple
layers simultaneously, you can call the view() method which will return a ViewStruct that
contains public fields referencing each layer [^0].

[jgm]

match ref_$fieldname() or mut_$fieldname().

This is surely a case where inline backticks should be used:

match ref_$fieldname() or mut_$fieldname().

So I'm not too swayed by this example. There are many cases (not just with math) where code symbols may have uninterpreted markdown interpretations if you don't put them in code backticks.